RNA-tailor: accurate gene-level identification of transcript isoform diversity from long reads

Abstract

Accurate splicing isoform identification is an essential need for progress of modern medicine and biological knowledge. The advent of long-read sequencing technologies opened up the possibility of sequencing full length transcripts. Most of the existing methods aims to predict isoforms at genome scale. However, solving all isoforms at genome scale is not always needed, mostly when one is interested in a single gene or a few genes. It also brings algorithmic constraints, encouraging the use of less accurate alignment algorithms and downgrading sensitivity of methods to determine the repertoire of isoforms. The aim of RNA-tailor is to make available an easy-to-use tool to perform single gene resolution of alternative splicing isoform repertoire with high accuracy. To leverage such precision, RNA-tailor uses a combination of exact alignment algorithm and context aware alignment corrections. By analyzing both real and simulated datasets, we show that RNA-tailor is able to achieve higher sensitivity.